Scene Classification on RESISC-45 (TR=10%)
94.73AccuracySelectiveMAE
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SelectiveMAEBackbone=ViT-L [2], Params (M)=307, Throughput/Minute=533k2024.06 | 94.73 | — | |
| SelectiveMAEBackbone=ViT-L [2], Params (M)=307, Throughput/Minute=533k, Pre-training=OpticalRS-13 (4M)2024.06 | 94.57 | — | |
| RSP-VITAE#P=19.3M, FLOPS=119.1G2024.03 | 94.41 | — | |
| LSKNet-S#P=14.4M, FLOPS=54.4G2024.03 | 94.27 | — | |
| RingMoBackbone=Swin-B [102], Params (M)=882024.06 | 94.25 | — | |
| ConvNext#P=28.0M, FLOPS=93.7G2024.03 | 94.07 | — | |
| LSKNet-T#P=4.3M, FLOPS=19.2G2024.03 | 94.07 | — | |
| RSP-Swin#P=27.5M, FLOPS=37.7G2024.03 | 94.02 | — | |
| RVSABackbone=VIT-B+RVSA [21], Params (M)=862024.06 | 93.93 | — | |
| RSP-R50#P=25.6M, FLOPS=86.3G2024.03 | 93.93 | — | |
| RVSA#P=114.4M, FLOPS=301.3G2024.03 | 93.92 | — | |
| CTNet2024.03 | 93.9 | — | |
| SelectiveMAEBackbone=ViT-B [2], Params (M)=86, Throughput/Minute=556k2024.06 | 93.7 | — | |
| SelectiveMAEBackbone=ViT-B [2], Params (M)=86, Throughput/Minute=556k, Pre-training=OpticalRS-13 (4M)2024.06 | 93.35 | — | |
| GRMANet#P=54.1M, FLOPS=171.4G2024.03 | 93.19 | — | |
| KFBNet2024.03 | 93.08 | — | |
| AGOSBackbone=DenseNet-121, Type=CNN2022.05 | 93.04 | — | |
| FSCNet#P=28.8M, FLOPS=166.1G2024.03 | 93.03 | — | |
| FENet#P=23.9M, FLOPS=92.0G2024.03 | 92.91 | — | |
| AGOSBackbone=ResNet-101, Type=CNN2022.05 | 92.91 | — | |
| F2BRBM#P=25.6M, FLOPS=86.3G2024.03 | 92.74 | — | |
| GFMBackbone=Swin-B [102], Params (M)=882024.06 | 92.73 | — | |
| CADNetType=CNN2022.05 | 92.7 | — | |
| ScaleMAEBackbone=ViT-L [2], Params (M)=307, Throughput/Minute=206k2024.06 | 92.63 | — | |
| MBENet#P=23.9M, FLOPS=108.5G2024.03 | 92.5 | — | |
| AGOSBackbone=ResNet-50, Type=CNN2022.05 | 92.47 | — | |
| MAEBackbone=ViT-B [2], Params (M)=86, Throughput/Minute=264k, Pre-training=OpticalRS-13 (4M)2024.06 | 92.44 | — | |
| MBLANet2024.03 | 92.32 | — | |
| MBLANetType=CNN2022.05 | 92.32 | — | |
| LSENet#P=25.9M, FLOPS=>86.3G2024.03 | 92.23 | — | |
| SatLasBackbone=Swin-B [102], Params (M)=88, Throughput/Minute=243k2024.06 | 92.16 | — | |
| UPetu#P=87.7M, FLOPS=>322.2G2024.03 | 92.13 | — | |
| LSENetType=CNN2022.05 | 91.93 | — | |
| DMSMILType=CNN2022.05 | 91.93 | — | |
| EAM#P=>42.3M, FLOPS=>164.32024.03 | 91.91 | — | |
| SatMAEBackbone=ViT-L [2], Params (M)=307, Throughput/Minute=205k2024.06 | 91.72 | — | |
| MSDFFType=CNN2022.05 | 91.56 | — | |
| IDCCP#P=25.6M, FLOPS=86.3G2024.03 | 91.55 | — | |
| TOVBackbone=ResNet-50 [101], Params (M)=262024.06 | 90.97 | — | |
| ViT-B#P=86.0M, FLOPS=118.9G2024.03 | 90.96 | — | |
| GASSLBackbone=ResNet-50 [101], Params (M)=262024.06 | 90.86 | — | |
| MG-CAP#P=>42.3M, FLOPS=>164.3G2024.03 | 90.83 | — | |
| MG-CAPType=CNN2022.05 | 90.83 | — | |
| MSANet#P=>42.3M, FLOPS=>164.32024.03 | 90.38 | — | |
| LiGNetType=CNN2022.05 | 90.23 | — | |
| MF2NetType=CNN2022.05 | 90.17 | — | |
| SeCoBackbone=ResNet-50 [101], Params (M)=262024.06 | 89.64 | — | |
| SCCov#P=13.0M2024.03 | 89.3 | — | |
| DCNNType=CNN2022.05 | 89.22 | — | |
| CACOBackbone=ResNet-50 [101], Params (M)=262024.06 | 88.28 | — | |
| MS2APType=CNN2022.05 | 87.91 | — | |
| SSL4EOBackbone=ViT-S [2], Params (M)=222024.06 | 87.6 | — | |
| Contourlet CNNType=CNN2022.05 | 85.93 | — | |
| RANetType=CNN2022.05 | 85.72 | — | |
| MIDCNetType=CNN2022.05 | 85.59 | — | |
| MSCPType=CNN2022.05 | 85.33 | — | |
| SPPNetType=CNN2022.05 | 82.13 | — | |
| AlexNetType=CNN2022.05 | 76.69 | — | |
| VGGNet-16Type=CNN2022.05 | 76.47 | — | |
| GoogLeNetType=CNN2022.05 | 76.19 | — | |
| AGANType=GAN2022.05 | 72.21 | — | |
| MARTAType=GAN2022.05 | 68.63 | — | |
| BoVW(SIFT)Type=Hand-crafted2022.05 | 41.72 | — | |
| Foundation ModelGeneral pre-training=-, Remote sensing pre-training=0.28M, Parameters=32M (Swin-T)2026.05 | — | 82.71 | |
| RingMo-LiteGeneral pre-training=-, Remote sensing pre-training=0.15M, Parameters=30M (Swin-T)2026.05 | — | 89.85 |